Predicting early vs late vs no re-exacerbation in COPD patients:Analysis of ECLIPSE cohort

Some patients may experience clustering of re-exacerbations following an index exacerbation event. We explored predictors of early and late re-exacerbations in the 180-days period from an exacerbation event. The ECLIPSE study included COPD patients(GOLD II-IV) recruited from outpatient clinics. We analysed data from patients who experienced ≥1 moderate/severe exacerbation during the 3-year study and had complete data for key variables. Exacerbations were classified as moderate(treated with OCS and/or antibiotics) or severe(requiring hospitalisation). Following the resolution of their index exacerbation, a new event was categorized as an ‘early’(within 90 days) or ‘late’(91-180 days) re-exacerbation and patients who did not re-exacerbate within 180 days were called ‘non’ re-exacerbators. A multivariate partial proportional odds model to identify factors predictive of ‘early’ vs ‘late’ vs no re-exacerbation was built using 10-fold cross validation. Of 1420 patients with ≥180 days follow-up after their index exacerbation, 951(67%) experienced ≥1 re-exacerbation. Lower FEV1 prior to index and higher baseline SGRQ score were associated with a higher risk of early re-exacerbation vs experiencing late or no re-exacerbation and a higher risk of any re-exacerbations (early or late) vs no-re-exacerbation[Table]. Prior frequent exacerbation history and season were also significantly associated with a higher risk of early and any re-exacerbation. Among patients with COPD, early re-exacerbation(within 90 days) and any re-exacerbation(within 180 days), are both significantly associated with a history of frequent severe exacerbations, poorer lung function and health status, and index AECOPD in the Winter season.